Electro-oxidation kinetics of adsorbed CO on platinum electrocatalysts
Introduction
The electro-oxidation of carbon monoxide from platinum surfaces has long been studied as a benchmark for activity of these catalysts toward carbonaceous adsorbates. For fuel cells, this is important from both the standpoint of CO tolerant catalysts that are resistant to poisoning by impurities in the fuel stream, and for fuel cells that operate on liquid fuels, where CO is generally recognized as the primary intermediate in the electro-oxidation of carbonaceous fuels (Steele and Heinzel, 2001; Liu et al., 2006; Lamy et al., 2001; Arico et al., 2001; Beden and Lamy, 1998). Methanol electro-oxidation (Ross, 1991; Lu et al., 2000; Tripkovic et al., 2002; Seland et al., 2006) in particular has been studied extensively in an attempt to improve electrocatalysts for this application. However, a complete fundamental understanding of how COads behaves on platinum surfaces remains elusive.
Often, voltammetric techniques are applied to study the characteristic CO oxidation from the surface in electrolyte with and without the presence of a bulk-CO source. This source can be from dissolved CO (g), an alcohol, or other carbonaceous fuel in the electrolyte. The characteristic size and shape of cyclic voltammetry (CV) peaks for platinum can be used as a fingerprint for surface sites and electro-oxidative activity.
The hydrogen-region characteristic current peaks, which result from H-adsorption (cathodic) and desorption (anodic) during low potential sweeps on platinum, are known from single crystal studies to be associated with the location or site of the hydrogen adsorbate (Markovic and Ross, 2002a; Love et al., 1986; Kita et al., 1990; Markovic et al., 1991, Markovic et al., 1997a; Lopez-Cudero et al., 2003; Teliska et al., 2004). These peaks do not translate in a straightforward manner to nanoparticle platinum surfaces because the disordered nanoparticle electrodes are unlikely to be composed of extended planar faces. The general position and shape of the peaks, however, is attributed to a family of energetically similar sites on the platinum surface. For polycrystalline platinum, two major peaks are typically observed in the hydrogen desorption region, with a lower potential peak comprising hydrogen from disordered (1 1 0)-types of sites, and a higher potential peak associated with hydrogen from disordered (1 0 0)-like types of sites, with both peaks containing mixed contributions from (1 1 1)-types of sites. These peaks are referred to as weakly and strongly adsorbed hydrogen on platinum (Markovic et al., 1991, Markovic et al., 1997a; Lopez-Cudero et al., 2003; Teliska et al., 2004).
In our previous work (McGrath et al., 2007), we also grouped the sites on Pt nanoparticles into two categories, weakly bound (WB) and strongly bound (SB), associated with the low- and high- potential hydrogen desorption peaks, respectively, from the voltammogram. The WB and SB designation simply refers to the family of sites associated with Hads. In that work, we observed fundamental differences in the hydrogen region of the platinum voltammogram resulting from different preparations of COads on Pt/C. These surface layers were prepared from partial adsorptions or partial oxidations of adsorbates created from CO (g) bubbled through electrolyte, or electrochemically adsorbed CH3OH. We found that the evolution of the hydrogen region resulting from various surface preparations is a result of the different activities of WB and SB sites toward CO (g) or CH3OH adsorption and COads oxidation. This was in agreement with single crystal studies in the past that have reported trends in the ability of different surface sites to adsorb species from the bulk (Bittins-Cattaneo et al., 1988), and the activity of different surface sites toward COads oxidation (Markovic et al., 1997b, Markovic et al., 1999a, Markovic et al., 1999b).
In this work, we extend the voltammetric analysis to the behavior of the COads oxidation peak resulting from various surface preparations on Pt black. Because the complete oxidation of the CO intermediate is recognized as the slowest step in the mechanism, it is possible to prepare a surface layer from the accumulation of COads. Surfaces are prepared through partial adsorption or partial oxidation of COads from CO (g) or CH3OH in 0.5 M H2SO4. Careful replacement of the CO or CH3OH-containing electrolyte with clean electrolyte allows one to study the size and shape of the peak associated with the oxidation of this prepared surface layer, without obfuscation from the adsorption and oxidation of dissolved bulk reactants.
We observe differences in the total COads oxidation current (which appears to be composed of two separate COads oxidation peaks) related to the preparation of the COads on the surface. A modified Butler–Volmer model is developed to probe the separate behavior of these two peaks resulting from different surface preparations. This analysis affords insight into the kinetic parameters governing the COads oxidation reactions. We parse the corresponding hydrogen region into WB and SB families of sites, as in our previous work (McGrath et al., 2007). Changes in WB- and SB-site occupation are correlated to the separate CO oxidation peaks. Limitations of this modeling scheme are discussed.
Section snippets
Experimental apparatus and procedures
Cyclic voltammetry (CV) experiments were performed in a modified 50 mL 3-neck flask that allowed for liquid flow from a 1 L reservoir of clean, 0.5 M H2SO4 (Certified ACS Plus Grade, Fisher Scientific with DI-water filtered in a Millipore Organ-X system, 18 MΩ). The cell compartment and the reservoir were degassed with UHP Argon gas (Praxair). The working electrode consisted of a Pt-black layer bonded with Nafion to a 0.5 cm2 glass slide coated with a 1 μm gold film. The catalysts used in this work
Results
Fig. 1 shows voltammograms (obtained at a sweep speed of 25 mV/s) for a clean Pt surface and for the same electrode after exposure to CO (g). The missing area in the hydrogen region (below 400 mV) for the CO-treated electrode is used to calculate the number of sites covered by COads. The total charge of COads oxidation is calculated from the area under the dashed curve (CO-covered) after correction for the double-layer charging current and background currents shown by the solid curve (clean
Discussion
Trends in the surface coverage and Neps values can lead to valuable insight about surface species resulting from different preparations on platinum black. We recognize that a platinum surface saturated with COads does not correspond to a true monolayer (Rush et al., 2001; Rush, 1998) equivalent to the amount of adsorbate that one can obtain from Hads in acidic media. With corrections for double layer effects and concurrent adsorption of anions from sulfuric acid, however, we can use our
Conclusions
Electroanalytical modeling the COads oxidation peak has yielded some insight into the behavior of adsorbates on the platinum black surface. Previous work prompted us to parse the total oxidation current into contributions from two different families of Pt surface sites, with each family of sites resulting in an oxidation peak. These peaks correspond well with the behavior on the individual families of sites (WB and SB) that are labeled based on the potential at which hydrogen desorbs from them.
Acknowledgements
This material is based upon work supported by the US Army Research Laboratory and the US Army Research Office under contract/Grant no. 48713CH.
References (45)
- et al.
Electrochim. Acta
(1988) - et al.
Electrochimica. Acta
(2003) Surf. Sci.
(2002)- et al.
Surf. Sci.
(1998) - et al.
Electrochim. Acta
(1977) - et al.
Electrochim. Acta
(2007) - et al.
Electrochim. Acta
(2007) - et al.
J. Electroanal. Chem.
(2002) - et al.
JEC
(1990) - et al.
Electrochim. Acta
(2008)